10 
NEIL E. STEVENS 
killing frost in autumn is also available for most of the stations and the 
regular Weather Bureau observation stations give the percentage of 
possible sunshine each day as well as atmospheric pressure and direc- 
tion and velocity of wind. 
Among the climatic elements recorded, any direct relation between 
atmospheric pressure and growth is very difficult to trace. Wind 
velocity and light while undoubtedly important for a green plant 
probably have little relation to the growth of Endothia parasitica, 
especially since the advancing edge of the mycelium is under the 
unbroken bark of the host tree. A careful study of the Weather 
Bureau data shows no correlation between amount of growth and 
either the prevailing direction of the wind or the number of clear 
days during the period investigated. The writer's laboratory experi- 
ments also have failed to demonstrate any relation between the amount 
of light and the growth and fructification of the fungus even when 
growing on the surface of culture media. 
Precipitation 
As regards precipitation, there are plainly two elements to be 
considered: the amount of rainfall and its distribution. Table III 
gives the monthly precipitation for each station during the course 
of the investigations, Table IV the number of days with more than 
.01 inch precipitation for each year during the same period. Careful 
examination of rainfall data fails to show any relation between either 
amount or frequency of rain and the amount of growth of the fungus. 
Amherst, Mass., had practically the same rainfall as Charlottesville, 
Va. Yet the growth at the latter point averaged nearly four inches 
greater. Williamstown, Mass., on the other hand, had a much 
smaller rainfall than Amherst but showed about the same amount of 
growth. Amherst, Mass., and Hartford, Conn., had much greater 
rainfall for the year ending in August than for that ending in May, 
yet the amount of growth was practically the same. 
While the different localities show considerable variation both in 
the amount of rainfall and in the number of days with rain this seems 
to have no relation to the amount of growth. This is probably best 
shown by the curves (figs, i and 2) of rainfall and number of days with 
precipitation. The various points of these curves are expressed in 
percentage of the rainfall and number of days with rain at Charlottes- 
ville, Va. The irregularity of the rainfall curves as compared with 
